The prediction accuracy of the Polymarket platform averages 94.1% four hours before market closure, according to a dashboard created by data processing specialist Alex McCullough on Dune.
A month before the event resolution, the accuracy remains at a high level, around 90%.
The expert also noted a clear bias in Polymarket’s markets. He attributed this to a tendency towards consensus, herd instinct, and low liquidity.
In an interview for the platform’s blog, McCullough stated that investors typically overestimate the probability of an event by one or two points.
Long-term forecasts show greater accuracy, the expert clarified. This is because they generally include unlikely elements. For instance, the platform listed 17 candidates for the US presidential election, including rapper Kanye West. However, it was evident that only Donald Trump and Kamala Harris had a real chance of winning.
For this reason, sports event markets like “one-on-one” show less precision, around 80%. However, their accuracy steadily increases as the event approaches.
McCullough admitted that his research was hindered by the lack of substantial historical data on Polymarket. In some segments he wished to analyze, the data volume “looked simply ridiculous.”
According to Polymarket Analytics, the volume of bets on the US presidential election reached a record on the platform, nearing $3.7 billion. Sports is a rapidly growing sector, with the NBA finals (~$1.6 billion), NFL (~$1.2 billion), UEFA Champions League (~$1 billion), and English Premier League (~$0.8 billion) leading the way.
Earlier, Forbes included Polymarket founder Shayne Coplan in its “30 Under 30” list for 2025 in the “Finance” category.